The complement system is an important component of the innate immune defense. A prerequisite for the complement system to exert its function is its activation, which can occur through three different ways: the classical, the lectin and the alternative pathways. Invading pathogenic microorganisms (e.g. bacteria, viruses, and fungi) can directly initiate each distinct pathway before the adaptive immune response is developed.
The complement cascade, however, if inappropriately activated, can cause a significant amount of inflammation, tissue damage, and other disease states such as the autoimmune diseases. Disease states implicating the complement system in inflammation and tissue damage include the following: the intestinal inflammation of Crohn's disease (Ahrenstedt et al., 1990), thermal injury (burns, frostbite) (Gelfand et al, 1982; Demling et al., 1989), hemodialysis (Deppisch et al., 1990; Kojima et al., 1989), and post pump syndrome in cardiopulmonary bypass (Chenoweth et al., 1981; Chenoweth et al., 1986; Salama et al., 1988), supposedly it is involved in the development of fatal complication in sepsis (Hack et al., 1989) and causes tissue injury in animal models of autoimmune diseases. The complement system is also involved in hyperacute allograft and hyperacute xenograft rejection (Knechtle et al., 1985); Guttman, 1974); Adachi et al., 1987). Complement activation during immunotherapy with recombinant IL-2 appears to cause the severe toxicity and side effects observed from IL-2 treatment (Thijs et al., 1990). Further deleterious effects of improper activation or overactivation of the complement system is described e.g. in U.S. application Ser. No. 2002037915.
Based on increased incidence of infections in individuals with MBL deficiency, there are indications in the art that the lectin pathway is associated with the following diseases: HIV (increased susceptibility of infection), cystic fibrosis, systemic lupus erythematosus, rheumatoid arthritis, recurrent miscarriage, meningitis, cryptospirodiosis, chronic hepatitis C (Dumestre-Perard, 2002) (as a disease modulator). Complement activation, contributing to the inflammatory reaction upon observed in reperfusion injury is mediated through the lectin pathway (Monsinjon, 2001, Collard, 2000). In a rat animal model, blockade of the lectin pathway protected the heart from ischemia-reperfusion by reducing neutrophil infiltration and attenuating proinflammatory gene expression. (Jordan, 2001)
It is of particular importance therefore to study key molecules of the complement system, their structure and function, therefore to obtain these molecules or functional or folded fragments thereof in sufficient quantities and in a pure form to obtain appropriate research tools, to develop assays for detecting said molecules and to find, design or raise molecules for effecting the function of the complement system or for supplying deficiencies of it and also treating decease conditions associated with irregular working of this system.
The activation of the complement system (like of other proteolytic cascades) results in the sequential activation of serine protease zymogens. The first step in the lectin and the classical pathways is the binding of a specific recognition molecule (MBL or C1q, respectively) to activator structures, which is followed by the activation of associated serine proteases (Gál, 2001).
Although the lectin pathway was discovered more than a decade ago (Kawasaki, 1989), there are many uncertainties concerning the composition of the activation complex and the substrate specificities of the MBL-associated serine proteases (MASPs). MBL is a member of the collectin family of proteins and binds to specific carbohydrate arrays on the surface of various pathogens through C-type lectin domains (Turner, 1996). Up to date three MBL-associated serine proteases have been described. First, a single enzyme ‘MASP’ was identified and characterized as the enzyme, which is responsible for the initiation of the complement cascade (i.e. cleaving C2, C4 and possibly C3) (Matsushita, 1992/Ji, Y-H., 1993). Later it turned out that ‘MASP’ is in fact a mixture of two proteases: MASP-1 and MASP-2 (Thiel, 1997). It was demonstrated, that the MBL-MASP-2 complex alone is sufficient for complement activation (Vorup-Jensen, 2000). This is a significant difference from the C1 complex, where the coordinated action of two serine proteases (C1r and C1s) leads to the activation of the complement system. Here, C1q is the recognition subunit of the complex, while C1r and C1s are highly specific serine proteases (with Mrs 86.5 kDa and 80 kDa, respectively), which are responsible for the catalytic function of C1. A specific feature of the C1r and C1s serine proteases is that they form a distinct structural unit, the Ca2+-dependent C1s-C1r-C1r-C1s tetramer, which makes possible the coordinated action of the two enzymes within the C1 complex. This tetramer associates with C1q to yield the heteropentameric C1 complex. C1r and C1s are present in the C1 complex in zymogen form, and become activated after C1q binds to an activator. The first enzymatic event during the activation process is the autoactivation of C1r. Activated C1r then activates zymogen C1s, which in turn cleaves C4 and C2.
The role of MASP-1 in the MBL-MASPs complex remained unknown. It was proposed, that MASP-1 could directly cleave C3 and thereby activate complement (Matsushita, 1995/Matsushita, 2000), but other laboratories debated these results (Wong, 1999/Zhang, 1998). Recently, a novel protease, MASP-3 has been isolated, however, its function is yet to be resolved (Dahl, 2001). Several lines of evidences suggest that there are different MBL-MASPs complexes (Thielens, 2001/Dahl, 2001) and a large fraction of the total MASPs in serum is not complexed with MBL (Terai, 1997/Thiel, 2000).
The MASPs together with C1r and C1s, form a family of proteases with identical domain organization (Sim, 2000/Volanakis, 1998). In these enzymes the first N-terminal CUB domain is followed by an EGF-like domain and the second CUB domain. A tandem repeat of complement control protein modules (CCP1 and CCP2) precedes the C-terminal serine protease domain (SP). Upon activation an Arg-Ile bond is cleaved in the serine protease domain of these zymogens.
Although the substrate specificities of MASP-1 and MASP-2 has been studied using natural and recombinant proteins, several important questions remained unanswered in the art. MASP-1 was shown to cleave C3 and C2 (Matsushita, 2000), but this action may not be sufficient for direct complement activation (Rossi, 2001). If the cleavage of C3 by MASP-1 proves to be insignificant then the field is still open to assess the biological importance of MASP-1. This could be possibly accomplished by identifying the range of its substrate specificity and the degree of its specific activity and by finding a ‘better’ natural substrate than C3. Previous studies showed that MASP-2 digested C2 and C4 efficiently, with rates similar to C1s, a classical pathway enzyme (Vorup-Jensen, 2000/Rossi, 2001). However, the contribution of the individual domains to the enzymatic properties of MASP-2 has not yet been determined. It is accepted that C1-inhibitor reacts with both proteases (Matsushita, 2000), but the rates of the reactions are unknown and the role of another inhibitory protein, alpha-2-macroglobulin, is rather debated (Wong, 1999/Rossi, 2001/Gulati, 2002). To sum up, differences in function and biological role of MASP-1 and MASP-2 are unclear according to the art.
Similarly, before the creation of the present invention, the role of the CCP and SP domains of C1r was unclear. The CCP repeat is about 60 residues in length and is widespread among complement proteins. It appeared to be likely that the CCP domains significantly contribute to the specificity of the interaction and catalytic properties of the γB fragment. For example a recent structural model of (γB)2 suggest a loose head-to-tail assembly of the monomers, where the γ-chain (the two CCP modules and the activation peptide) of one monomer interacts with the serine protease module of the other monomer (8).
Moreover, the autoactivation mechanism was unclear. Which domain is necessary to autoactivation? At all, are the individual domains separate folding units? Can folded fragments be prepared? Are they active?
The answer to these questions and other problems of the art is an important prerequisite of further research aiming the treatment of complement related disorders.
The main reason for these uncertainties concerning biological function of these proteases is the lack of availability of active and/or native recombinant proteins in a sufficient quantity and purity. The functional characterization has been retarded by the fact that their serum concentration is very low (in the case of MASPs: [MASP-1]=6 μg/ml, [MASP-2]≈2 μg/ml) (Hajela, 2002) which rendered their isolation extremely difficult. Most MASP preparations obtained from serum were usually cross-contaminated with other MASP species. Therefore, precise experiments could not be carried out in many cases. Also, though methods for determining in vivo levels of a mixture of MASPs have been known (e.g. U.S. Pat. No. 6,235,494 and U.S. Pat. No. 6,294,024) assessing MASP-1 and MASP-2 levels, differentiating one from the other, caused problems according to the art.
Though JP 7238100 (Matsushita et al, 1995) is directed to a monoclonal antibody against human MASP, the defined character of such an antibody is questionable since the publication does not make a distinction between MASP-1 and MASP-2. In general, due to the lack of sufficiently pure preparations antibodies for native MASP proteins could not be produced according to the art with a good reliability.
Similarly, a search for drugs, e.g. inhibitors, alleviating symptoms associated with overactivity of the complement system had been greatly hindered by the lack of availability of folded, possibly active fragments of key multidomain serine proteases of the complement system.
Also, in lack of a reliable and effective system for the recombinant preparation of fragments of complement serine proteases associated with recognition molecules of the complement system mutation studies and genetic engineering of them were difficult and the outcome was difficult to interpret.
Last but not least methods of the art for producing MASP proteins and fragments were relatively expensive and cumbersome: cheaper and more effective methods were needed.
There have been many attempts in the art to prepare recombinant MASP-1 and MASP-2. However, recombinant expression of the full-length MASPs provided serious difficulties. In WO 02/6460 (Jensenius and Thiel, 2002) cloning and sequencing of MASP-2 is described. However, the protein was not recombinantly prepared. Vorup-Jensen et al. transiently expressed human MASPs in HEK293 cells, but their recombinant MASP-1 had unexpected molecular mass and showed no enzymatic activity (Vorup-Jensen, 2000). Though Vorup-Jensen et al. succeeded in preparing an active MASP-2 protein, this protein carried a His tag. Moreover, the preparation process was prolonged and complicated. Chen et. al. tried to produce rat MASP-1 and MASP-2 in CHO cells, but the wild type proteases were cytotoxic to the cells and therefore only inactive mutants could be produced (Chen, 2001). Rossi et. al. expressed full-length human MASP-1 and MASP-2 in a baculovirus insect cell system, but due to the very low yield the proteases could not be purified to homogeneity (Rossi, 2001).
Rossi et al. (2001) expressed also CCP1-CCP2-SP fragments of MASP-1 and MASP-2 in a baculovirus insect cell system, which showed comparable enzymatic activities with the full-length molecules towards protein and ester substrates. Said fragments were secreted by the insect cells. However, CCP-SP fragment of MASP-2 could not be successfully produced in the baculovirus expression system (Rossi et al., 2001). Based on these results the authors concluded that the first CCP domain (CCP1) is crucial for activity and propose that the smallest active fragment of MASP proteins is the CCP1-CCP2-SP fragment. Moreover, expression in a baculovirus/insect cell system has obviously several disadvantages from the economic point of view that is relatively low yield, high costs and the complexity of purification process. Moreover proteins secreted from the insect cells are subject to a possible protease attack.
Previously, the baculovirus-insect cell system was used to produce recombinant C1r and C1s and their fragments (14, 20, 21). The yield of the secreted recombinant proteins, however, was found to be low. The catalytic C-terminal γB fragment of C1r, consisting of the two CCP domains followed by the activation peptide of the protease and the serine protease domain (B-chain), can be obtained by autolysis or by limited proteolysis of extrinsic proteases (e.g. thermolysin) (6). However, no recombinant production of the γB fragment of C1r or C1s had been disclosed before creation of the present invention. Also, no smaller recombinant fragments of the C1r or C1s catalytic region had been disclosed. In particular, no prokaryotic expression of C1r or C1s fragments had been suggested in the art.
To circumvent the problems of the art outlined above the present inventors decided to attempt recombinant expression of the catalytic fragments of multidomain serine proteases capable of binding to recognition molecules of the complement cascade. In spite of the fact that these are complex, multidomain proteins which are of mammalian origin, inventors decided to choose a prokaryotic, in particular a bacterial system, more particularly an E. coli based system.
To the best of their knowledge, the present Inventors were the first to express and successfully refold multidomain serine proteases in bacterial hosts.
It has long been accepted in the art that the proteolytic γB fragments of C1r and C1s, which consist of three domains: two CCP modules and the serine protease domain, retain the catalytic activity of the entire molecule both in terms of substrate specificity and catalytic efficiency (Villiers, 1985/Arlaud, 1986/Lacroix, 1989). Recent studies with recombinant fragments of C1r and C1s reinforced this view (Rossi, 1998/Lacroix, 2001). Since the MASP proteases share the same domain organization with C1r and C1s, it seems plausible that the CCP1-CCP2-SP fragments mediate the catalytic activity of these enzymes, as well. This is supported by the recent studies of Rossi et al. who showed comparable enzymatic activities with the full-length molecules towards protein and ester substrates and found that CCP1 domain is necessary to activity.
Applying a bacterial expression system, it is an object of the invention to produce functionally active fragments of MASP-1, MASP-2, MASP-3, C1s and C1r, with a yield sufficient for structural and functional characterization, as well as to provide a teaching to prepare the corresponding fragments of C1s and MASP-3. Furthermore, a further object of the invention is to create fragments from which the CCP domains preceding the SP domain were successively deleted and to create the respective expression vectors. It is a further object of the invention to functionally characterize the fragments.
Though in some bacterial systems, mainly in those providing low expression levels, the proteins could be expressed in a soluble form, overexpression systems are advantageous. A usual problem of overexpressing mammalian proteins in bacterial hosts is that the unfolded foreign protein forms inclusion bodies and has to be subjected to renaturation, though this feature may turn out to be an advantage: the expressed proteins are protected against protease attack. Despite a great number of proteins successfully renatured, protein refolding remains a problem to be solved on a case-by-case basis [Rudolph and Lilie, (1996)]. It is to be mentioned here that, as a matter of course, successful attempts are published and failures usually not. Nevertheless, even in the recent past protein refolding was considered as an extremely difficult task. By now it is generally approved that whereas in vitro refolding of single domain proteins is likely to be successful, refolding of multidomain proteins remains a problem the solution of which is far from being obvious (Fischer et al EPA 0 393 725 A1, Ambrosius et al., EPA 0 500 108). Furthermore, it is also well-known that protein folding is usually started at the N-terminal of the polypeptide chain. Therefore, if an N-terminal part of a protein is deleted, its refolding is significantly encumbered. Moreover, to the best of Inventors' knowledge, multidomain serine proteases, in particular of human origin, have not been prepared in a folded form in a prokaryotic expression system.
Being aware of the fact that larger molecules are more difficult to be refolded, but hoping that the presence of the N-terminal portion of the molecule may help inducing the folding process, Inventors attempted to renature the entire MASP molecules. After the failure of methods at hand, Inventors used a variety of additives and removed them in a stepwise manner which, of course, raises costs. In spite of this, no unambiguously positive results were obtained.
Surprisingly, inventors found that the C-terminal CCP1-CCP2-SP fragments (also named as γB-fragment after the nomenclature used for C1s and C1r proteins) of MASP proteins and of C1r and C1s proteins could be renatured with an improved renaturation method at a sufficient or, under preferred conditions, at a high yield.
Applying this method for smaller fragments success or promising results were achieved. In view of former results of their own and of those of Rossi et al (2001) it is also surprising that inventors found the MASP-2 CCP2-SP and the SP fragments to be active.
The inventors also recognized that a particularly improved method can be carried out if a temperature below 10° C. and a pH above pH 8.7, preferably above pH 9 or 10, is applied and, preferably, in the refolding buffer at least arginine is applied as a chaotropic agent.
In a similar expression system and renaturation method (which was slightly modified by applying higher temperature and a lower pH) inventors could express and isolate recombinant fragments of C1r as well.
Having now large amounts of pure fragments available, Inventors could provide a detailed functional characterization of the proteins. In particular, Inventors found differences in substrate and inhibitor specifities of MASP-1 and MASP-2, providing a basis for differentially measuring their level in serum or in a biological sample.
The fragments obtained according to the invention can be used advantageously e.g. in drug screening methods and for antibody production. Results also suggest possible applicability of them, in particular MASP-1, and of their inhibitors in medical treatments.